Re: [Gretl-users] RES: Same model, different position of variables, different results
Am 04.10.2012 20:53, schrieb Bruno Thiago Tomio:
Sorry, I did not get it, Pindar. Could you please be more specific?
I know just enough Portuguese (and also know gretl's messages...) to see
that you have an exact collinearity problem, so gretl drops some
arbitrary variables, and this arbitrary choice depends on the ordering.
Or is there more to it? If so, please show us something without the rank
deficiency.
cheers,
sven